Student Economic Review Vol.

XXXIII

An Analysis of the
Meltzer-Richard-Roberts
Assumption

Navika Mehta, Senior Sophister

Rising inequality is an issue that has plagued modern society for
generations. In this paper, Navika Mehta provides a comprehensive
critical analysis of the Meltzer – Richard – Roberts Assumption, that
rising inequality will lead to greater redistribution of social spending.
This has been widely used to explain the rise in social spending in the
20th century. However, the paper argues that voting behaviours, elite
influence, concepts of social mobility and fairness, and the irrational
human nature disproves the real-world application of the assumption.
In doing so she shows that social spending is impacted by a myriad of
factors that cannot simply be reduced to rising inequalities.

Introduction
I n an interview with the Rolling Stone Magazine on his book The Price of In-
equality, Nobel Laureate economist, Joseph Stiglitz explained that “High levels
of economic inequality leads to imbalances in political power as those at the top
use their economic weight to shape our politics in ways that give them more eco-
nomic power.” (Bernstein, 2012). He asserts that in the United States, outcomes of
political processes seldom reflect the interests of citizens. The problem of ‘asym-
metric information’ and disillusioned citizens is also highlighted as a contributing
factor that enables the use of economics for political agendas. Having said that,
equality is a pillar of democracy but inequality is a fact of life. Democracies help
in providing equal opportunities to all citizens; including equality in the form of
universal suffrage, in front of the law and through the absence of discrimination

64
 Economic Policy

based on race, religions, sexual orientation, gender, class etc. Essentially, democ-
racies across the world have seen rising inequalities and in particular, income
inequality. In short, the rich are getting richer and the poor, poorer. Through
recessions and depressions, social movements like the Occupy Wallstreet have
attempted to highlight this inequality with slogans like – “we are the 99%”.
That said, the idea of inequality and demand for support from the state is a
contemporary issue. Meltzer, Richard and Roberts give the assumption that ris-
ing inequalities will lead to greater redistribution or social spending (Meltzer and
Richard, 1981 & Roberts, 1977). This essay attempts to refute this assumption
and show that although this makes sense in theory, in reality, it does not apply.
It begins with explaining the Meltzer-Richard-Roberts assumption in detail and
evaluates how it is used to explain the rise in social spending during the 20th cen-
tury. It then gives three main arguments against this theory. Firstly, it highlights
the importance of turnout in an election and elite control influencing who the
median voter is and thus the outcome of the elections. Secondly, it shows that the
idea that economic growth will lead to increased redistribution because of high
inequalities does not apply in the developing world. Finally, it evaluates the role
of the concept of social mobility and individual beliefs about ‘fairness’ that go be-
yond economic interests and influence people’s choices. Hence proving that the
Meltzer-Richard-Roberts hypothesis is, essentially, flawed in reality.
The Meltzer-Richard-Roberts Assumption
A highly influential prediction about income inequality and redistri-
bution, based on the Median Voter Theorem was given by Meltzer and Richard.
The model, known as the Meltzer-Richard Model emphasizes how elections
are important in ensuring social and economic equality (Meltzer & Richard,
1981). They use the Median Voter Theorem, which shows that with universal
suffrage and majority rule, the median voter is the decisive voter (Roberts,
1977). They also use studies regarding the distribution of income to show that
the income distribution is skewed to the right, i.e., the mean income lies above
the median income, in order to prove their argument. The assumptions are that
there is a unidimensional policy space, that is redistribution, and all citizens are
paying a flat tax rate that is used to finance this redistribution. Moreover, all
citizens receive the same tax transfer and they are voting for their preferred tax
rate. The implication of their hypothesis is that the distance between the mean
income and the income of the decisive voter (median voter) is what determines
the size of the government. With a skewed income distribution as mentioned
above, there are more people earning income lower than the mean income,

65
 Student Economic Review Vol. XXXIII

which incentivises greater redistribution of income financed by incomes of

higher earning citizens. This implies that higher taxes and redistribution lead to
a reduction in the incentive to work in the first place, thereby lowering income
and increasing inequality (Meltzer & Richard, 1981).
They set out a static model, that individuals chose between consumption
and leisure and this choice is dependent on the tax rate and the size of transfer
payments. This tax rate and the size transfer payments further depend on the
voting rule and income distribution. Essentially, their argument follows two main
assertions. First, income inequality determines the amount of redistribution. Sec-
ond, the greater the inequality or distance between the median and mean income,
the greater will be the redistribution or social spending. This implies that in a gen-
eral equilibrium model, the government has only two functions of redistribution
and taxation (Meltzer & Richard, 1981). If the budget is balanced and voters are
fully informed, the decisive median voters’ choice is what determines the size of
the government. They explain that voters earning below the median voter will
prefer higher taxes and greater redistribution while those earning above the me-
dian voter will prefer the opposite. Thus, as mean income rises relative to the me-
dian voter income, taxes will also rise. Roberts gives an argument that is similar
in essence. He explains that the marginal tax rate chosen by majority voting will
be the highest possible due to the fact that if the median voter earns less than the
voter with the mean income and income distribution is skewed, then people with
low income will prefer a high tax rate (Roberts, 1977). The overall argument is
used to explain the rise in social spending in the 20th century.
An Explanation for the Rise in Social Spending
Meltzer and Richard use their argument to explain the rise of social spend-
ing in the 19th and 20th century with the rationale that as the number of voters
earning below the mean income increased, it shifted the median voter position and
increased redistribution and taxation. They also connect this to economic growth
leading to greater inequality and therefore to greater redistribution (Meltzer &
Richard, 1981). In the introduction of their paper, they mention that the share
of redistribution and income tax has increased in western European countries in
North America, and specifically in the US and Britain and that this rate of tax pro-
portional to income has existed for over a century. However, they do not claim
that this theory is limited to these countries, in their view, it is largely universal.
That said, the rise in social spending in the 20th century mainly occurred in the
democratic countries of Western Europe and North America. Meltzer and Rich-
ard use the spread of the franchise to argue that in the 19th and 20th century,

66
 Economic Policy

this led to an increase in voters with low income and it sh ifted the median voter
down to a lower income bracket and so taxes rose and as a result, social spend-
ing increased. Lindert argues that Meltzer and Richard fail to explain two facts
about this rise. First, the differences in the share of GDP between countries is not
explained. Second, there is ambiguity whether countries with higher redistribu-
tion and low incomes will glow slower than countries with lower redistribution
and high incomes (Lindert, 1996). He offers ‘competing hypothesis’ that give
alternative explanations for the rise in social spending and its eventual decline.
He presents the argument that the dispersion of income below the median voter
matters in determining redistribution. He highlights the influence of socialism,
socialist democratic parties and labour union as a factor that led to the rise in so-
cial spending in the 20th century. Moreover, he argues that the age distribution of
voters can also influence the government’s redistribution policies, as the elderly
are more likely to be in favour of contributions to health and pension. He uses the
deadweight costs theories to argue that high social spending in the long run and
the rising costs of increasing the size of government will eventually limit social
spending (Lindert, 1996). This is because the rising deadweight costs will reach
a point where it will eventually stop income growth. Dispersion of income is an-
other factor that influences social spending. The increase in inequality below the
median voter will lead to greater redistribution. After testing these hypotheses
on a sample of 19 countries between 1960 and 1981, Lindert finds that although
the Meltzer-Richard model seems theoretically plausible, there is little empirical
evidence for it having an impact in the 20th century. He explains that growth in
social spending is impacted by various factors like the diminishing age or income
effect. He predicts an end to this growth in the 21st century (Lindert, 1996).
Critique of Meltzer- Richard – Roberts Assumption
Voting Behaviour and Turnout
The following section gives three further arguments to counter the Melt-
zer-Richard-Roberts Model. Firstly, the importance of turnout and the reality of
elite control influencing voting decisions and outcomes. An important argument
countering the claims made by the Meltzer- Richard and Roberts Model is given
by Larcinese (2007). He explains two opposite effects of the rise in inequali-
ty. First, increased inequality would lead to the median voters’ preference for
greater redistribution. Second, increased inequality would mean differences in
turnout, wherein the rich and privileged citizens are more likely to vote than
those living in extreme poverty. This could be due to illiteracy, work restrictions,
lack of interest or awareness. Thus, the overall impact is that inequality would not

67
 Student Economic Review Vol. XXXIII

actually lead to more redistribution (Larcinese, 2007). He argues that although

it seems theoretically logical to say that countries that are poor and have greater
inequality are not growing due to excessive social spending, it is misleading in
reality. Meltzer and Richard do highlight the extension of voting rights as lead-
ing to an increase in voters below the median voter, however, they do not take
into account that the turnout is not 100% (Meltzer & Richard, 1981). Larcinese
studies 41 countries between 1972 and 1998 and does not find a clear correlation
between redistribution and inequality; explaining that country-specific factors
are influencing this relationship (Larcinese, 2007).
Elite Control of Democracy
This argument is supporting Acemoglu and Robinsons’ idea of ‘captured
democracy’, which describes elite control of the political system after democ-
ratization. The elites use their economic power to invest in their de facto power,
they influence parties through lobbying and control political ideology through
the media. Acemoglu and Robinson explained that the elites must contribute to
the collective de facto power that must be larger than the de jure power of poor
voters (Acemoglu & Robinson, 2008). For example, in a study on Clientelism in
India, Anderson et al. (2015) find that in the state of Maharashtra in areas where
Maratha landlords have power in the form of landholdings, democratic decisions
are pro-landlords and opposed to the poor because of their elite clientelist ties to
political parties. Furthermore, Luebker (2014) conducts an empirical analysis of
110 observations from 26 countries and found no relationship between inequal-
ity and redistribution. He explains that factors such as unemployment rates, an
increase in an ageing population are more effective in explaining the rise in social
spending. Further critiquing, that the Meltzer-Richard-Roberts model uses ratio-
nal choice and portrays humans as rational which is indifferent to the argument
in behavioural economics that humans take into account society, shared values,
ideas of social justice and fairness that further influence their economic decisions
(ibid).
Evidence from Developing Countries
Secondly, the Meltzer-Richard-Roberts model has been critiqued widely
in relation to the developing world. Keefer and Khemani (2005) argue that the
Meltzer-Richard-Roberts model does not hold in low income countries. Here,
the median voter is extremely poor and evidently, social services are lacking.They
highlight three market imperfection causing this outcome – “lack of information
about the performance of politicians, social fragmentation and identity-based vot-
ing, and absence of credibility of promises made by politicians” (ibid). They pro-

68
 Economic Policy

vide evidence that in developing countries, governments tend to prefer investing

in infrastructure and government jobs rather than social services. For example,
in India, when disadvantaged groups comprise of the vote bank more emphasis is
laid on increasing access to government jobs and welfare transfers to gain support
in the short term rather than investing in education and healthcare, which would
not cost as much and would be beneficial in the long term (ibid). However, if
politicians are concerned with only winning elections, short-term appeasement
is the chosen strategy.
Affinity-Based Voting and the Right to Vote
Moreover, Keefer and Khemani (2005) state that countries like India and
Nigeria experience ‘affinity-based voting’ which means that people are more
likely to vote on caste, religious, tribal or racial lines. This is not evidenced in the
United States or Western Europe. In order to explain their critique, they refer to
a study conducted to examine the disparity in outcomes of health and education
in two states in India – Kerala and Uttar Pradesh (UP). Between the 1960s and
1995, in each decade the average per capita public expenditure spending in Kera-
la was more than two times that of UP. UP allocated a large amount of resources
to state administration and very little to health and education while Kerala did the
opposite. They argue that these differences are due to the market imperfections
mentioned above. Moreover, citizens in Kerala are better informed and parties
compete on offering social services. UP, on the other hand, relies on clientelist
influence. High literacy rates in Kerala and relatively lower caste-based discrim-
ination compared to UP affected the outcome. Essentially, the Communist Party
in Kerala influenced and increased party competition for the dominant Indian
National Congress Party (INC), while in UP, the INC did not have a strong com-
petitor. In UP social services remain weak and people tend to vote on caste-based
lines. Thus, in a democracy, several different forms of electoral competition can
impact the outcomes of redistribution. Keefer and Khemani (2005) conclude
that although poor voters are highly active, the credibility of politicians in provid-
ing basic social services without pursuing clientelist agendas is needed to ensure
greater redistribution. Ahuja and Chhibber (2012) validate this by questioning
the reasons for the large turnout of the poor in elections in India. They conduct a
series of focus groups and open-ended interviews to find that although the poor
are not the ones gaining from elections, they see voting as a right more so than
a civic duty or a way to get economic benefits. They find that voting is an affir-
mation of citizenship (Ahuja & Chhibber, 2012). Thus, there are several factors
influencing turnout and income inequalities in developing countries. Even with

69
 Student Economic Review Vol. XXXIII

high turnout, social spending is still very low.

The concept of Social mobility
Finally, another criticism of the Meltzer-Richard-Roberts Model comes
from the perspective of political sociology in the form of the concept of social
mobility and fairness. People with income below the average, do still hope that
in future they will move up on the income scale. In a democracy, this is a valid
possibility. This idea of upward mobility is constructed in the form of expectation
that in the future their children may be the high-income earners and will lose
from high taxes. Benabou and Ok reflect (2001) on the “prospect of upward mo-
bility” (POUM) hypothesis. If citizens have low-risk aversion and they expect to
have higher income in future, then they would be less inclined to vote for higher
social spending. Alesina and Angeletos (2005) go a step further and question the
idea of the statement – “people get what they deserve and deserve what they
get”. The idea of fairness in the level of distribution is questioned. They focus on
the differences between Europe and the United States. While European countries
tend to have more effective redistribution policy, possibly due to the history of
class struggle, in the United States, the perception is that the wealthy have earned
their place. They also talk about the racial struggle in the US after the end of
slavery, they highlight that the reason for “poor deserves to be poor” idea is that
the median income voter is white and the low-income voter is black (Alesina &
Angeletos, 2005).
Fairness and irrational human behaviour
Piketty (1995) further explains the social mobility argument in shaping po-
litical attitudes and electoral outcomes. The argument follows that people have
varied experiences and they learn and believe differently about how taxation im-
pacts the society. He argues that it is important to look at these different beliefs
and ideas about the role of the government in the economy (Piketty, 1995). The
exposure to different information based on their socio-economic background
can highly impact the outcome, even if they are fully informed. It is important
to go beyond the economic interests of the citizens (Piketty, 1995). A study con-
ducted in the Slums of New Delhi by Banerjee et al. (2011) offers evidence for
this argument. The level of awareness and information that voters have are limit-
ed. They divided areas into treatment and control groups. The treatment groups
were given newspapers along with report cards that gave information about the
performance of the legislators running for elections (Banerjee, et al., 2011). Pub-
lic readings of the report card were conducted in the treatment areas. They found
that providing this information influenced the perception of the benefits from

70
 Economic Policy

voting and led to changes in turnout relative to the performance of the legislator.
Conclusion
In conclusion, the claim that greater inequality will lead to greater redis-
tribution because the Median voter will have a low income and so will vote for
higher taxes is not convincing.This assumption is flawed on several grounds.Voter
turnout is not taken into consideration, which is especially important as for the
majority of the 20th century, several low-income groups and women did not
have voting rights. Today, in developing countries, even if the poor are coming
out to vote in large numbers, social services are still lacking. Elites control the
economic power used for campaigning in elections, they influence voters through
the media, use lobbying and clientelism to influence the government. Developing
countries have large income inequalities that are not addressed with large social
spending, even if there is a high turnout. Governments are more interested in
spending on infrastructure than education and health. Moreover, Identity-based
voting is rampant. Factors including social mobility or expectations of having
higher income in future prevent people from voting for politicians who promise
higher taxation and greater redistribution. The culture of a country and percep-
tions of ‘fairness’ influence citizens voting decisions. Finally, human beings must
not be looked at as rational profit maximising agents, because they typically are
not. Humans are influenced by the social fabrics of their society and the infor-
mation that form their beliefs and ideologies. These highly influence their voting
behaviour. Hence as Stiglitz explained, political and economic power go hand in
hand and till the time a few influential high-income earners exist in a capitalist
system, social spending can never be a priority for governments.

71
 Student Economic Review Vol. XXXIII

Reference List:
1. Acemoglu, D. & Robinson, J. A., 2008. Persistence of Power, Elites,
and Institutions. American Economic Review, March, 98(1), pp. 267-
293.
2. Ahuja, A. & Chhibber, P., 2012. Why the Poor Vote in India: “If I Don’t
Vote, I Am Dead to the State”. Studies in Comparative International
Development, June, 47(4), pp. 389-410.
3. Alesina, A. & Angeletos, G.-M., 2005. Fairness and Redistribution.
American Economic Review, September, 95(4), pp. 960-980.
4. Anderson, S., Francois, P. & Kotwal, A., 2015. Clientelism in Indian
Villages. American Economic Review, June, 105(6), pp. 1780-1816.
5. Banerjee, A. V., Kumar, S., Pande, R. & Su, F., 2011. Do Informed Vot-
ers Make Better Choices? Experimental Evidence from Urban India.
Unpublished Manuscript , November .
6. Benabou, R. & Ok, E. A., 2001. Social Mobility and the Demand for
Redistribution: The Poum Hypothesis. The Quaterly Journal of Eco-
nomics, May, 116(2), pp. 447-487.
7. Bernstein, J., 2012. The Price of Inequality: Interview with Joseph E.
Stiglitz. [Online]
8. Available at: https://www.rollingstone.com/politics/politics-news/
the-price-of-inequality-interview-with-joseph-e-stiglitz-51549/
9. [Accessed 24 November 2018].
10. Keefer, P. & Khemani, S., 2005. Democracy, Public Expenditures, and
the Poor: Understanding Political Incentives for Providing Public Ser-
vices. The World Bank Research Observer, 20(1), pp. 1-27.
11. Larcinese, V., 2007. Voting over Redistribution and the Size of the Wel-
fare State: The Role of Turnout. Political Studies, October , 55(3), pp.
568-585.
12. Lindert, P. H., 1996. What Limits Social Spending?. Explorations in
Economic History, January , 33(1), pp. 1-34.
13. Luebker, M., 2014. Income Inequality, Redistribution and Poverty:
Contrasting Rational Choice and Behavioural Perspectives. Review of
Income and Wealth, March, 60(1), pp. 133-154.

72
 Economic Policy

14. Meltzer, A. H. & Richard, S. F., 1981. A Rational Theory of the Size
of Government. Journal of Political Economy, October , 89(5), pp.
914-927.
15. Piketty, T., 1995. Social Mobility and Redistributive Politics. The Qua-
terly Journal of Economics, August, 110(3), pp. 551-584.
16. Roberts, K. W., 1977. Voting over Income Tax Schedules. Journal of
Public Economics, December , 8(3), pp. 329-340.

73